Synthesis, Photophysical, Electrochemical, Tumor-Imaging, andPhototherapeutic Properties of Purpurinimide-N-substitutedCyanine Dyes Joined with Variable Lengths of LinkersMichael P. A. Williams,† Manivannan Ethirajan,† Kei Ohkubo,§ Ping Chen,∥ Paula Pera,† Janet Morgan,‡

William H. White, III,† Masayuki Shibata,*,⊥ Shunichi Fukuzumi,*,§,# Karl M. Kadish,*,∥

and Ravindra K. Pandey*,†

†PDT Center, Department of Cell Stress Biology, Roswell Park Cancer Institute, Buffalo New York 14263, United States‡Department of Dermatology, Roswell Park Cancer Institute, Buffalo New York 14263, United States§Department of Material and Life Sciences, Graduate School of Engineering, Osaka University, ALCA, Japan Science and TechnologyAgency (JST), Osaka, Japan∥Department of Chemistry, University of Houston, Houston, Texas 77204-5003, United States⊥Department of Health Informatics, University of Medicine & Dentistry of New Jersey−SHRP, Newark, New Jersey 07107, UnitedStates#Department of Bioinspired Science, Ewha Womans University, Seoul 120-750, Korea

*S Supporting Information

ABSTRACT: Purpurinimide methyl esters, bearing variablelengths of N-substitutions, were conjugated individually to acyanine dye with a carboxylic acid functionality. The resultsobtained from in vitro and in vivo studies showed a significantimpact of the linkers joining the phototherapeutic andfluorescence imaging moieties. The photosensitizer−fluorophoreconjugate with a PEG linker showed the highest uptake in theliver, whereas the conjugate linked with two carbon units showedexcellent tumor-imaging and PDT efficacy at 24 h postinjection.Whole body imaging and biodistribution studies at variable timepoints portrayed enhanced fluorescent uptake of the conjugatesin the tumor compared to that in the skin. Interestingly, theconjugate with the shortest linker and the one joining with twocarbon units showed faster clearance from normal organs, e.g., the liver, kidney, spleen, and lung, compared to that in tumors.Both imaging and PDT efficacy of the conjugates were performed in BALB/c mice bearing Colon26 tumors. Compared to theothers, the short linker conjugate showed poor tumor fluorescent properties and as a corollary does not exhibit the dualfunctionality of the photosensitizer−fluorophore conjugate. For this reason, it was not evaluated for in vivo PDT efficacy.However, in Colon26 tumor cells (in vitro), the short linker was highly effective. Among the conjugates with variable linkers, therate of energy transfer from the purpurinimide moiety to the cyanine moiety increased with deceasing linker length, as examinedby femtosecond laser flash photolysis measurements. No electron transfer from the purpurinimide moiety to the singlet excitedstate of the cyanine moiety or from the singlet excited state of the cyanine moiety to the purpurinimide moiety occurred asindicated by a comparison of transient absorption spectra with spectra of the one-electron oxidized and one-electron reducedspecies of the conjugate obtained by spectroelectrochemical measurements.

1. INTRODUCTION

Current treatment modalities for patients afflicted by cancerinclude surgery, radiation therapy, chemotherapy, and arelatively novel option called photodynamic therapy(PDT).1−5 Surgery is used to excise abnormal growths andsurrounding tissue, but the procedure is invasive and may becomplicated by relapse of the cancer if not all of the tumor cellsare removed. Chemotherapy employs different cytotoxicchemicals to attack or block specific cellular and molecularmechanisms that aid tumor growth. Unfortunately, patients on

chemotherapeutics suffer from side effects due to adverse drugtoxicities. Radiation uses ionizing energy to attack neoplasticcells, but it is nonspecific and may cause damage to surroundinghealthy tissue, which can lead to the occurrence of secondarycancers. PDT uses a drug known as a photosensitizer, light, andoxygen to destroy tumors and their surrounding vasculature.

Received: June 30, 2011Revised: September 14, 2011

Article

pubs.acs.org/bc

© XXXX American Chemical Society A dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXX

PDT has several advantages in that (i) there is no systemic,organ, or tissue toxicity, (ii) it is noninvasive, and (iii) it can beused repeatedly as a primary or adjuvant treatment.6−9 With thecurrent advent of imaging technologies to monitor tumorresponses to treatment, we have shown that the utilization of acertain photosensitizer−cyanine dye conjugate which containsthe tumor-avid photosensitizer 3-(1′-hexyloxyethyl)-pyropheo-phorbide-a (HPPH) for treatment and a fluorophore cyanine

dye for optical imaging could be highly advantageous to treatdeeply seated tumors and monitor the treatment response.10

The ability to image real time events using fluorescence hasmade optical imaging an attractive modality to study cellularand molecular events11 and to visualize events in vivo, especiallyin tumors.12 Noninvasive in nature, optical imaging instrumentsare simpler and less expensive to operate and can allow preciseassessment of the location and size of a tumor, providing

Scheme 1. Synthesis of Purpurinimide−Cyanine Dye Conjugates Joined with Variable Lengths of Linkersa

aReagents and conditions: (a) Zn(OAc)2, reflux in MeOH, 2 h; (b) Pd/carbon, H2 THF, 12 h; (c) TFA, 2 h, RT; (d) NH2−NH2(anhydrous),pyridine DCM/HCL, RT; (e) R1(CH2)2, DCM, CH2N2, KOH/MeOH; (f) R2(C2H4O)2C2H4, CH2N2, KOH/MeOH; (g) DCM/TFA (1:4), 2h; (h) DMTMM, DMF, 12 h.

Scheme 2. Preparation of a Cyanine Dye with Carboxylic Acid Functionality

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXB

information on its invasiveness in adjacent tissue.13,14 The workdiscussed herein describes the synthesis, molecular modeling,photophysical, electrochemical, tumor-imaging, and the photo-therapeutic potential of a series of longer wavelengthphotosensitizer (purpurinimides, 700 nm) joined to a cyaninedye with variable lengths of linkers. Among the near-infrared(NIR) dyes, cyanine dyes in general have shown enormouspotential for optical imaging.15−19

2. RESULTS AND DISCUSSION2.1. Chemistry. For the synthesis of purpurinimide-N-

substituted cyanine dye conjugates, purpurin-18 methyl ester 1was synthesized from chlorophyll-a by following the literatureprocedure, which was converted into mesopurpurin-18 methylester 2 by following the standard methodology.20 Reaction of 2with hydrazine hydrate gave N-amino mesopurpurin-18 methylester 3 in 50% yield.21 By following a similar approach, 1 wasindividually reacted with diethyl diamine and carbamic acid toproduce the corresponding N-substituted analogues 4 and 5both with 89% yield. Reaction of these N-substituted purpurin-18 methyl esters with a cyanine dye 6 containing a carboxylicacid functionality (obtained in 80% yield by reactingcommercially available IR820 with p-thiol-benzoic acid) gavepurpurinimide−cyanine dye conjugates 7−9 in which twochromophores (photosensitizer and fluorophore) are joined atvariable linker lengths with yields ranging from 32 to 37%. Thereaction sequences for the preparation of the conjugates andthe cyanine dyes are shown in Schemes 1 and 2.Structures of the intermediates and the final products were

confirmed by 1H NMR and CHN/mass spectrometry(including HRMS) analysis. The purity of the final conjugates(compounds 7−9) was determined by HPLC analysis.2.2. Molecular Modeling. In order to examine the effects

of linker lengths joining the two chromophores (conjugates 7,8, and 9) in photophysical, electrochemical, tumor-imaging andphotodynamic efficacy, a molecular modeling study wasperformed. The three-dimensional model of the conjugateswere built with the molecular modeling software SYBYL(Tripos Inc., St. Louis, MO) followed by the energyminimization with the PM3 semiempirical molecular orbitaltheory using the Spartan 02 software package (Wave functionInc., Irvine, CA). Figure 1 shows the resulting structures of the

conjugates 7, 8, and 9 (Scheme 1) in extended conformation.

The distances between the nitrogen atom in the purpurinimide

ring system and the sulfur atom adjacent to the cyanine dye of

the conjugates were calculated and are displayed in Figure 1. Asexpected, the smaller the linker length, the shorter the observeddistance between two chromophores.Since the above structures are appropriate for single

molecules in gas phase, they may not be relevant to thestructures of the conjugates in solution, in cells, or in vivoenvironments where physical properties and biological activitiesof these conjugates are examined. In order to gain someinsights, the conformational flexibility of these conjugates wereexamined with the stochastic search module of MOE softwareusing the MMFFs force field and charges (ChemicalComputing Group, Montreal, Quebec, Canada). Again, thedistance between two chromophores was measured for some ofthe resulting low energy conformers as shown in Figure 2.As can be seen from the Figure 2 caption, the distances

between the two chromophores remain more or less constantfor conjugates 7 and 8, while it shows a large variation forconjugate 9. It is reasonable considering the length of the linkerfor these conjugates. Although the mean distances are similarbetween conjugates 7 and 8, closer inspections reveal thesignificant difference between these conjugates. First, thestandard deviation is much smaller for conjugate 7 than for8, reflecting the limited flexibility of 7. Second, this limitedflexibility is also the source of the limited range in the relativeorientation of two chromophores for conjugate 7 as shown inFigure 2a where various low energy conformers of conjugate 7are superimposed using the purpurinimide ring as a reference.It clearly shows that the cyanine dye and the linker can assumemany different conformations but that they cannot assume aconformation that brings two chromophores close together.Thus, the purpurinimide ring remains exposed to itssurroundings for this conjugate. Compared to conjugate 7,some part of the purpurinimide ring is covered by the cyaninedye or thiophenol group in conjugate 8, as shown in Figure 2b.Although there are a large variety of conformers for conjugate9, there are some conformers where a large portion of thepurpurinimide ring is covered by the cyanine dye and/orthiophenol group as shown in Figure 2c. Photophysicalexperiments clearly indicated that intramolecular energytransfer occurred from the singlet excited state of thepurpurinimide moiety to the cyanine moiety with the rate oftransfer increasing with decreasing linker distance. This is anopposite trend from the linker length dependence on the invitro PDT efficacy. The photophysical results are consistentwith the distance dependence of energy transfer, while theconformational flexibility may explain the excellent in vitro PDTefficacy of conjugate 7 and the decrease of efficacy withincreasing linker length. If the purpurinimide rings are notexposed to the environment, the compound may not be able toproduce singlet oxygen because it cannot interact with watermolecules. Alternatively, even if singlet oxygen is produced, itmay not reach to the cellular target because the singlet oxygenproduced may interact with the closely seated cyanine dyebefore it reaches the target site(s).2.3. Photodynamics. In order to examine the photo-

dynamics of purpurinimide, cyanine, and the conjugates, time-resolved transient absorption spectra were recorded byfemtosecond laser flash photolysis in deaerated DMSOsolutions as shown in Figure 3. Transient absorption bandsof purpurinimide 4 observed at 500 and 530 nm taken at 2 psafter femtosecond laser excitation (Figure 3a) are assigned asthe singlet excited state of 4 (14*). The transient absorption of14* decreased as the absorption band of the triplet excited state

Figure 1. Three-dimensional representation of the conjugates withdifferent lengths between the cyanine dye and the purpurinimide inextended conformation obtained from molecular modeling and energyoptimization with PM3.

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXC

of 4 appeared. The transient absorption band of the singletexcited state of cyanine 6 is observed at 586 nm as shown inFigure 3b. The lifetimes of the singlet excited state of 4 and 6were determined from decays of the absorption at >3 ns and500 ps, respectively. In the case of conjugate 7 shown in Figure3c, the transient absorption bands due to both singlet excitedstates of purpurinimide and cyanine were observed at 500, 530,and 586 nm at 2 ps because both moieties were excited by laserflash excitation at 410 nm. The absorption bands at 500 and530 nm due to the singlet excited state of the purpurinimidemoiety decreased, accompanied by an increase in theabsorption band at 586 nm due to the singlet excited state ofthe cyanine moiety. This indicates that energy transfer from thesinglet excited state of purpurinimide to the cyanine moietyoccurs efficiently to afford the singlet excited state of cyanine.The energy transfer dynamics in 7 were determined from the

rise of the absorption band at 586 nm as shown in the left panelof Figure 4a. The energy transfer rate constant was determinedto be 6.7 × 1011 s−1. Similarly, the energy transfer rate constantsof conjugates 8 and 9 were also determined from the rise of theabsorption band at 586 nm in Figure 4b and c to be 1.4 × 1011

and 1.3 × 1011 s−1, respectively. The energy transfer rateconstant increases with decreasing distance between the energydonor and acceptor estimated from the theoretical calculations(vide supra). The decay time profiles at 586 nm are shown onthe right side of Figure 4. The decay rate constants of 7−9agree well with the value of the cyanine reference 6 due tointersystem crossing to the triplet excited state. An energytransfer from the singlet excited state of the purpurinimidemoiety to the cyanine moiety is feasible because the singletenergy of the purpurinimide (1.76 eV) is higher than that of thecyanine (1.41 eV), which were obtained from the absorptionand fluorescence maxima, λ abs = 689 nm and λ fl = 720 nm for4, and λ abs = 849 nm and λ fl = 915 nm for 6, respectively. Thus,in the purpurinimide−cyanine systems, intramolecular energytransfer at the singlet excited state occurs efficiently from thepurpurinimide moiety to the cyanine moiety, followed byintersystem crossing to the triplet excited state of the cyaninemoiety. No electron transfer from the purpurinimide moiety tothe cyanine moiety was observed as expected from thedisfavored energetics as indicated by the redox potentials ofthe conjugates (vide inf ra).2.4. Electrochemical Properties. Cyclic voltamograms of

7−9 are illustrated in Figure 5, which also includes thereference compounds 4 and 6. The first one-electron oxidationpotentials of 7−9 (Eox) are all located at E1/2 = 0.52 V vs SCE,which is similar to the Eox value of compound 6 (0.56 V vsSCE). The first one-electron reduction potentials of 7−9(Ered1) are also identical at E1/2 = −0.55 V vs SCE, which agreeswith the Ered1 value of cyanine 6 (−0.56 V vs SCE). The secondone-electron reduction potentials of 7−9 (Ered2) are also nearlythe same and range from −0.64 to −0.69 V vs SCE, whichagree with the Ered1 value of compound 4 (−0.69 V vs SCE).Thus, there seems to be little or no interaction between thepurpurinimide and cyanine moieties in the conjugates 7−9,irrespective of the difference in the distance between them.The second one-electron oxidation of the purpurinimide

moiety of the conjugates 8 and 9 is irreversible, and the peakpotential ranges from 0.89 to 0.92 V vs SCE, which is similar tothat of 4 (0.94 V). The energy of the charge-separated state of7−9 to be produced by electron transfer from thepurpurinimide moiety to the singlet excited state of the cyanine

Figure 2. (a) Various conformers for conjugate 7 with limited flexibility in the linker are superimposed using the purpurinimide ring as reference.Standard color for each atom type is used. The purpurinimide ring (shown in the left portion of this figure) is exposed to surroundings. (b) Examplesof the low energy conformer for conjugate 8. The purpurinimide moiety is shown in orange and the cyanine dye−thiophenol moiety in cyan, and thelinker regions are based on standard atom type based color. There are some overlaps between the purpurinimide ring and the cyanine dye ring. (c)Example of the low energy conformer for conjugate 9. The same color scheme as that in panel b is used. Because of the flexibility of the long linker,diverse conformations are possible, and some of the low energy conformers show an extensive interaction between the purpurinimide ring and thecyanine dye−thiophenol moiety. Mean distance between the two moieties in panel a (conjugate 7) = 8.18 Å (SD 0.10); in panel b (conjugate 8) =8.55 Å (SD 0.51); and in panel c (conjugate 9) = 9.85 Å (SD 2.27).

Figure 3. Transient absorption spectra of purpurinimide 4, cyanine 6,and conjugate 7 in deaerated DMSO taken at 2, 100, and 3000 ps afterfemtosecond laser flash excitation at 410 nm.

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXD

moiety is roughly estimated to be 1.6 eV, which is higher thanthe energy of the singlet excited state of cyanine (1.41 eV). Theenergy of the charge-separated state of 7−9 in the reversedirection from the singlet excited state of the cyanine moiety tothe purpurinimide moiety is also estimated to be 1.74 eV, whichis higher than the energy of the singlet excited state of cyanine(1.41 eV). Thus, electron transfer from the singlet excited stateof the cyanine moiety to the purpurinimide moiety isenergetically feasible. However, no electron transfer occurredfrom the purpurinimide moiety to the singlet excited state ofthe cyanine moiety as shown in Figure 3. This indicates that theintersystem crossing from the singlet excited state of thecyanine moiety to the triplet excited state is much faster thanthe electron transfer.The absence of a charge-separated state of 7−9 was further

confirmed by measuring the purpurinimide radical cation andthe cyanine radical anion by use of a spectroelectrochemistry(vide inf ra). Figure 6a shows the spectral changes (in blue)which occurred during the one-electron reduction of cyanine 6at an applied potential of −0.70 V. A new absorption maximum

at 531 nm, which is assigned to the cyanine radical anion,appears, and this was accompanied by a disappearance ofabsorption bands at 773 and 849 nm due to cyanine. A cleanisosbestic point is seen in this transfer. The radical anion can befurther reduced to the dianion at an applied potential of −1.60V vs SCE. A similar spectral change is observed for the firstone-electron reduction of conjugate 7 (Figure 6b), where theradial anion of the cyanine moiety has an absorption band at533 nm. Figure 6c shows the spectral changes upon the one-electron reduction of purpurinimide 4 at an applied potential of−0.90 V, in which the Soret band at 417 nm and the visibleband at 689 nm are decreased in intensity, while an obviousradical band appears at 593 nm. These spectral changes shownin red are quite similar to the second reduction of conjugate 7at −0.80 V also shown in red (see Figure 6b). A comparison ofthe spectral change for 6, 7, and 4 during each reductionillustrates the similarity between conjugates 7, cyanine 6, andpurpurinimide 4, also suggesting that the purpurinimide andcyanine moieties in 7 are almost independent of each otherduring electron transfer processes.

Figure 4. Rise and decay time profiles of the absorbance at 586 nm of (a) 7, (b) 8, and (c) 9 in deaerated DMSO after femtosecond laser excitationat 410 nm. Left panels: short time range. Right panels: long time range.

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXE

Figure 7a shows the spectral changes which occur upon theone-electron oxidation of cyanine 6 at an applied potential of0.70 V. A new absorption maximum appears at 650 nm, whichis assigned as due to the cyanine radical cation. The absorptionband due to the cyanine radical cation in conjugate 7 is slightlyblue-shifted to 637 nm (Figure 7b). The purpurinimide radicalcation could not be observed by spectroelectrochemistrybecause of irreversible oxidation, as shown in Figure 7c,where the absorption band due to 4 only disappears at anapplied potential of 1.0 V vs SCE. The absence of an absorptionband at 533 nm due to the radical anion of cyanine (Figure 6b)or the absorption band at 637 nm due to the radical cation ofthe cyanine (Figure 7b) confirms that there is no electrontransfer from the purpurinimide moiety to the singlet excitedstate of the cyanine moiety or from the singlet excited state ofthe cyanine moiety to the purpurinimide moiety in theconjugate.2.5. Biological Studies. 2.5.1. In Vitro PDT Efficacy.

Because of the insoluble nature of conjugates 7−9, variousformulations were used to dissolve the compounds inappropriate concentrations. Among all the formulations,reasonable solubility was obtained in both 1% Tween 80 in5% dextrose−water solution.22 The in vitro PDT efficacy ofconjugates 7−9 was determined in Colon26 cells by followingthe standard MTT assay23 (for details, see the ExperimentalProcedures). In brief, the cells were incubated for 24 h with 7,8, and 9 at variable concentrations and then exposed to laserlight (1.0 J/cm2) at 695, 713, and 710 nm (the longestabsorption wavelengths corresponding to the purpurinimideportion for the respective conjugates). The PDT efficacy (cell

Figure 5. Cyclic voltammograms of compounds 4 and 6−9 in DMSOcontaining 0.1 M TBAP at a scan rate of 0.1 V/s.

Figure 6. Thin-layer UV−visible spectra of (a) 6, (b) 7, and (c) 4 upon the controlled reducing potentials in DMSO containing 0.1 M TBAP. Theblue spectral changes in conjugate 7 correspond to a reduction on the cyanine dye moiety. The red spectral changes are assigned at the reduction atthe purpurinimide unit of the conjugate.

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXF

kill) was calculated from 4 replicate wells in 3 separateexperiments, and each standard error is representative of 3separate experiments. The results, summarized in Figure 8 and

Table 1, show that increasing the distance between thephotosensitizer and cyanine dye decreases the in vitro PDTefficacy. The LD50s of the conjugates were determined by thebest fit curves (plotted in Sigmaplot) to the dose−response

data. The IC50 values confirmed a decreased in vitro PDTefficacy in this series of conjugates by increasing the linkerlength between the purpurinimide and the cyanine dye.

2.5.2. In Vivo Fluorescence Imaging. Cyanine dye 6 has therequired photophysical properties (long wavelength absorptionwith significant Stokes shift), for fluorescence imaging but wasnot tumor-avid.24 To investigate the use of tumor-avidpurpurinimides as vehicles to deliver the nontumor-avidcyanine dye to the tumor, we used fluorescence imaging as atool to investigate the effect of the linker in vivo.25 Compared tothe therapeutic dose, the imaging dose was quite low, andtherefore, we were able to evaluate the imaging potential of thethree conjugates 7−9. Each conjugate was injected i.v. via thetail vein at a dose of 0.03 μmol/kg into BALB/c mice (9 mice/group) bearing Colon26 tumors on the shoulder.Figure 9 represents fluorescent images of the tumors at 24,

48, and 72 h postinjection of conjugates 7, 8, and 9 (3 mice/time point) after each drug was injected. The color scalerepresents fluorescence intensity, where blue is the lowestintensity, and red/white is the highest. After imaging, the miceat each time point were sacrificed, and the organs wereremoved and imaged ex vivo to demonstrate the distributionand clearance rates of the conjugates. Conjugate 8 exhibited thehighest fluorescent intensity in the tumors of all thecompounds at 24 h and was gradually clear from the tumorby 72 h. Conjugate 9 showed a moderate but lowerfluorescence intensity than compound 8, which cleared fromthe tumor at a much faster rate. Interestingly, conjugate 7 gavelow, barely noticeable fluorescence intensity in the tumor,which could either indicate poor fluorescence quantum yieldfor the cyanine dye portion of this compound or low tumorselectivity.The ex vivo fluorescence intensities of various organs (ear,

heart, kidney, liver, lung, muscle, small intestine, spleen,stomach, and tumors) are shown in Figure 10. As expected,conjugate 9 in which the photosensitizer is joined with thecyanine dye via a PEG linker exhibited highest fluorescence(high uptake) in the liver compared to 7 and 8. The highaffinity for this compound to the liver could be due to the more

Figure 7. Thin-layer UV−visible spectra of (a) 6, (b) 7, and (c) 3 upon the controlled reducing potentials in DMSO containing 0.1 M TBAP.

Figure 8. MTT plotted assay of 7, 8, and 9 incubated for 24 h atvarious concentrations with Colon26 cells followed by a light dose of1.0 J/cm2 at a dose rate of 3.2 mW/cm2.

Table 1. IC50 (50% Cell Kill) Concentrations of Conjugates7, 8, and 9a

conjugate IC50 (μM) distance (Å) between the two moieties

7 (short linker) 0.22 8.188 (medium linker) 0.32 8.559 (long linker) 0.69 9.85aAverage distance of various conformers is shown in the caption toFigure 2.

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXG

lipophilic nature of the linker connecting the purpurinimideand cyanine dye. Interestingly, the PEG linker also showedcomparable tumor avidity. Since the purpose of this study is tocreate a molecule that can act as both a fluorescent imaging anda therapeutic agent in vivo, conjugate 7 was not examined anyfurther due to its weak uptake in tumors compared to that ofthe other two conjugates, determined by fluorescence imaging.However, further studies at lower time points may produceimproved efficacy, and these studies with this and other agentsare currently in progress.

2.5.3. Comparative In Vivo Tumor Uptake of theConjugates. In vivo reflectance spectroscopy was used todetermine the maximum uptake of the conjugates at varioustime points.26 Mice were injected through the orbital venousplexus, the reflectance in tumors was measured at 24 and 48 h,

and the drug concentrations were determined (see Figure 11).This approach was extremely useful in providing the knowledgeof the pharmacokinetic characteristics (especially the clearancefrom the tumor) of the conjugates and also the shift in their invivo absorption. Under a similar dose of the drug (2.5 μmol/kg), conjugate 8 showed maximum tumor uptake (BALB/cmice bearing Colon26 tumors) at 24 h postinjection, whichdecreased over the next 24 h. Conjugate 9 showed a smallincrease in uptake at 48 h compared to that at 24 h.Comparatively, conjugate 8 showed a higher tumor uptake

than 9 at both 24 and 48 h. Since our objective has been todevelop a candidate with both imaging and therapy capabilitiesand conjugate 7 showed limited PDT efficacy, it was decidednot to explore its imaging potential.

2.5.4. In Vivo Photosensitizing Efficacy. Preliminary in vivoPDT efficacy of conjugates 8 and 9 was measured by tail veininjection into BALB/c mice (5 mice/group) bearing Colon26tumors (average tumor volume 27 mm3) at a dose of 2.5 μmol/kg (which gave approximately 50% tumor cure with theHPPH−cyanine dye conjugate, our lead compound).10 Thetumors were illuminated with light (135 J/cm2, 74 mW/cm2) at24 h postinjection, with the best time point for the maximumphotosensitizer uptake in the tumor determined by fluores-cence imaging and in vivo reflectance spectroscopy. From thepreliminary results summarized in Figure 12A, it can be seenthat compared to 8, the conjugate 9 exhibited limited long-termefficacy and that all mice showed tumor regrowth at 10−12days post-treatment. Under similar treatment parameters,conjugate 8 showed 100% tumor cure, and no tumor growthwas observed at day 60. We further investigated the efficacy ofconjugate 8 at variable doses (1.5, 2.0, and 2.5 μmol/kg), andthe light dose was kept the same (135 J/cm2, 75mW/cm2).From the results summarized in Figure 12B, it can be seen thatat the lowest dose of 1.5 μmol/kg, conjugate 8 showed limitedefficacy, but at higher doses, a significant PDT response was

Figure 9. Comparative whole body images of conjugates 7−9 along with cyanine dye 6 (dose: 0.03 μmol/kg) in BALB/c mice bearing Colon26tumors. Images were taken at 24, 48, and 72 h postinjection using a 12 bit Nuance camera (CRI, Worburn, MA) with a 782 continuous wave laserfor excitation and a 800 and 830 nm long pass filter to collect fluorescence. Images were then analyzed using Image J software.

Figure 10. Fluorescence biodistribution of conjugates 7−9 and variousorgans [3 mice (BALB/c bearing Colon26 tumors]. Organs from themice were removed and then imaged/analyzed at 24 h postinjection(dose: 0.03 μmol/kg). The fluorescence intensity of each organ wasanalyzed by Image J software.

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXH

observed (2.0 μmol/kg, 3/5 mice, and at 2.5 μmol/kg, 5/5mice, were tumor free on day 60).

3. CONCLUSIONS

An efficient synthetic pathway was developed to createpurpurinimide−cyanine dye dual functional agents. In vitroresults suggested that increasing the carbon chain length from ashort linker to a long linker decreases the PDT efficacy of thecompounds. Fluorescence imaging showed that 7 (short linker)had poor imaging capabilities compared to that of 8 (mediumlinker) and 9 (long linker) and was not studied further.Conjugate 9 exhibited the highest liver fluorescent values. Thismay be due to the more lipophilic nature of the linkerconnecting the two molecules, which shows higher affinity tothe liver. In vivo reflectance spectroscopy showed that conjugate8 exhibited higher tumor uptake than 9 and should thereforeproduce in vivo PDT efficacy. The comparative in vivo PDTconfirmed the higher efficacy of conjugate 8 over 9 andproduced 100% tumor response (5/5 BALB/c mice bearingColon26 tumors were tumor free after day 60 post-treatment),whereas 9, under similar treatment parameters showed limitedPDT activity. This study with a small number of conjugatesindicates that conjugate 8 with a medium length of linker showspotential for both tumor imaging and therapy. However, further

studies with a larger group of compounds should help inselecting the best candidates, and the synthesis of the relatedanalogues is underway. The low PDT and imaging capabilitiesof conjugate 7 could be due to its faster in vivo clearance.Therefore, further studies with conjugate 7 at shorter timeintervals between the injection of the drug and light treatmentare also in progress.

4. EXPERIMENTAL PROCEDURES

All chemicals were of reagent grade and used as such. Allreagents were purchased from Aldrich chemical company andwere used as received. All photophysical experiments werecarried out using spectroscopic grade solvents. Solvents weredried using standard methods unless stated otherwise.Reactions were carried out under nitrogen atmosphere andwere monitored by precoated (0.20 cm) silica TLC plasticsheet (20 cm × 20 cm) strips (POLYGRAM SIL N-HR) andor/UV−visible spectroscopy. UV−visible spectra were recordedon a Varian Cary 50 Bio UV−visible spectrophotometer usingdichloromethane/methanol as solvent unless otherwise speci-fied. 1H NMR spectra were recorded on Varian 400spectrometers at 303 K in CDCl3 or ∼10% of CD3OD inCDCl3 or DMSO-d6. Proton chemical shifts (δ) are reported inparts per million (ppm) relative to CDCl3 (7.26 ppm), CD3OD

Figure 11. (A) In vivo absorption spectra of conjugates 7 and 8 determined by in vivo reflectance spectroscopy in a Colon26 tumor implanted inBALB/c mice at 24 h postinjection. (B) Tumor vs skin uptake of the conjugates at 24 and 48 h postinjection (drug dose: 2.5 μmol/kg).

Figure 12. (A) Comparative in vivo photosensitizing efficacy of conjugates 8 and 9 in BALB/c mice (5 mice/group) bearing Colon26 tumors at adose of 2.5 μmol/kg, and (B) in vivo efficacy of 8 at variable drug doses. The tumors were exposed to light (135 J/cm2 and 75mW/cm2) at theirlongest wavelength absorption at 24 h postinjection (i.v.) of the photosensitizers. The tumor growth was measured daily. Under the treatmentparameters used, the mice treated with conjugate 8 at a dose of 2.5 μmol/kg did not show any tumor regrowth, and the tumors were flat on day 60post-treatment.

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXI

(3.34 ppm) or TMS (0.00 ppm). Coupling constants (J) arereported in Hertz (Hz), and s, d, t, q, p, m, and br refer tosinglet, doublet, triplet, quartet, pentet, multiplet, and broad,respectively. Mass spectral data (Electro Spray Ionization, ESI,by fusion) were obtained from Biopolymer Facility, RoswellPark Cancer Institute, Buffalo, NY, and HRMS data wereobtained from the Mass Spectrometry Facility, Michigan StateUniversity, East Lansing, MI. Elemental analysis were done atMidwest Microlab LLC., Indianapolis, IN.4.1. N-Amino Purpurinimide Methyl Ester (3).

Mesopurpurin-18 methyl ester (2) (0.138 mmol, 80 mg) wasdissolved in 15 mL of anhydrous pyridine and stirred underargon. Hydrazine anhydrate (1.3 mmol, 42 mg, 42 μL) wasadded to 2 mL of anhydrous pyridine, which was then addedslowly to the stirring mixture. The completion of the reactionwas checked by UV−vis in regular intervals until thewavelength intensity at λmax of 701 nm was maximal. After 4h, 1 N HCl (25 mL) and dichloromethane (20 mL) wereadded, and the stirring was continued for another 1.5 h. Thereaction mixture was then transferred into a separation funnel.Additional dichloromethane (200 mL) was added, the organiclayer was washed with water (200 mL × 3). The dichloro-methane (DCM) layer was separated, dried over anhydroussodium sulfate, and filtered.. The crude reaction productobtained after evaporating DCM was purified on an aluminacolumn by eluting with a 1% MeOH/DCM solvent system toafford a purple/magenta crystal of 321 (42.4 mg, 50% yield).UV−vis λmax (in CH2Cl2): 359 nm (ε = 5.8 × 104), 416 (ε =10.0 × 105), 548 (ε = 1.9 × 105), 643 (ε = 8.60 × 103), 700 (ε= 4.94 × 104) . 1H NMR (400 MHz, 3 mg/1 mL CDCl3, δppm) 9.07, 8.95, and 8.42 (each s, 1H for 5H, 10H and 20H),5.72 (broad s, N−NH2), 5.29 (m, 1H, for 17H), 4.33 (m, 1Hfor 18H), 3.63 (q, 2H for 81CH2) 3.62 (s, 3H for 12CH3), 3.61(s, 3H for 172CO2CH3), 3.33 (m, 2H for 31CH2) 3.18 (s, 3Hfor 2CH3), 3.00 (s, 3H for 7CH3), 2.78 and 1.978 (m, 2H for171 CH2), 2.46 (m, 2H for 172 CH2) 1.76 (d J = 7.2 Hz, 3H for18CH3), 1.66 (t, J = 7.2 Hz, 3H for 82CH3), 1.50 (m, 3H for32CH3). EIMS (m/z): 595 (M + H). Elemental Anal. Calcd. forC34H38N6O4: C, 68.67; H, 6.44; N, 14.13. Found: C, 68.70; H,6.43; N, 14.01.4.2. N-Boc Diethlyene Purpurinimide Methyl ester

(4a). Purpurin-18 methyl ester (50 mg, 0.0865 mmol) (1) wasadded to a dry 100 mL round bottomed flask and put underhouse vacuum for 20 min. N-Boc ethylenediamine (48.50 mg,0.30275 mmol) was placed under nitrogen for 10 min and thenadded to the flask containing Purpurin-18 using a long needle.Under a nitrogen atmosphere, 5−10 mL of anhydrous DCMwas added to the reaction mixture, which was stirred for 39 h,monitoring with UV−vis. A wavelength of 665 nm indicatedthe opening of the six-membered anhydride ring system. It wasthen treated with diazomethane, the intermediate amide, asmethyl ester was not isolated, and immediately treated with acatalytic amount of KOH/MeOH. The base catalyzed intra-molecular cyclization afforded the desired analogue exhibitingthe long wavelength absorption at 707 nm. The reactionmixture was purified by preparative plates using a 5% MeOH/DCM solvent system, and the desired band was scratched offand resuspended in 5% methanol/dichloromethane. The silicawas removed by filtration. Solvents were evaporated, and theresidue 4a was precipitated with DCM/n-hexane (60 mg, 86%yield). UV−vis λmax (in CH2Cl2): 365 nm, 418 nm, 550 nm,707 nm. 1H NMR (400 MHz, 3 mg/1 mL CDCl3, δ ppm) 9.40,9.20, and 8.56 (each s, 1H for 5H, 10H and 20H); 7.83 (dd, J =

11.6, 5.2 Hz 1H, 31CHCH2); 6.24 (d, J = 16.4 Hz, 1H, trans32CHCH2), 6.12, (d, J = 8.4 Hz, 1H cis 32CHCH2); 5.46(br s, 1H, N-H-CH2−CH2−N-Boc); 5.37 (d, J = 4 Hz, 1H for17H), 4.62 (m, 2H, NH−CH2−CH2−N-Boc); 4.37 (q, 1H for18H); 3.59 (s, 3H, 172CO2CH3); 3.49 (q, 2H for NH−CH2−CH2-N-Boc); 3.79 (d, J = 5.2 Hz, 1H for 81CH2); 3.73 (s, 3Hfor 12CH3) ; 3.33 (s, 3H for 2CH3); 3.04 (s, 3H for 7 CH3);2.72, 2.46, 2.44, and 2.06 (each m, 1H for 2 × 171H and 2 ×172H); 1.79 (d, J = 8 Hz, 3H for 181 CH3); 1.60 (t, J = 7.6 Hz3H for 82 CH3); −0.128 and −0.215 (br s, 1H for 2NH). EIMS(m/z): 721 (M + H).4.3. Diethlyene Amino Purpurinimide Methyl Ester

(4). N-Boc diethlyene purpurinimide methyl ester (4a) wastreated with a 4:1 ratio of TFA/dry DCM and stirred for 2 h.The TFA was removed by high vacuum for 5 h, and theremainder was purified on an alumina column with 5% MeOH/DCM as the solvent system. The resulting compound wasisolated as a dark purple liquid, compound 4, which wasconcentrated by evaporating the liquid under vacuum to afforddark purple crystals (54 mg, 89% yield) UV−vis λmax (inCH2Cl2): 365 nm (ε = 4.5 × 104), 419 nm (ε = 1.18 × 105),551 nm (ε = 4.6 × 104), 706 nm (ε = 4.2 × 104). 1H NMR(400 MHz, 3 mg/1 mL CDCl3, δ ppm) 9.65, 9.44, 8.78 (each s,1H for 5 H, 10H and 20H); 8.03 (dd, J = 17.6, 11.6 Hz, 1H,31CHCH2), 6.45 (d, J = 18.0, 1H, trans 32CHCH2), 6.34(d, J = 11.6, 1H cis 32CHCH2), 5.58 (d, J = 8.4 Hz, 1H for17H); 4.81 (t, J = 6.8 Hz, 2H for N−CH2-CH2−N), 4.58 (q, J= 7.2 Hz, 1H for 18H), 3.95 (s, 3H, 12CH3), 3.81 (s, 3H,172CO2CH3), 3.80 (q, J = 7.6 Hz, 2H, 81CH2), 3.54−3.59 (m,5H, 2CH3 and N−CH2-CH2-N), 3.27 (s, 3H, 7CH3), 2.96,2.67, 2.62, and 2.22 (each m, 1H, 2 × 171H and 2 × 172H),2.00 (d, J = 7.2 Hz, 3H, 18CH3), 1.82 (t, J = 7.6 Hz, 3H,82CH3) 0.08 and 0.00 (each br s, 1H, 2NH). EIMS (m/z):621.1 (M + H). Elemental Anal. Calcd. for C40H48N6O6: C,69.66; H, 6.50; N, 13.54. Found: C, 68.63; H, 6.51; N, 12.88.4.4. N-Boc Glycol Purpurinimide Methyl Ester (5a). N-

Boc-2,2′-(ethylene-1,2-dioxy) bisethylamine was prepared byfollowing the literature procedure.10 It (50 mg, 0.2015 mmol)was dissolved in 5 mL of dry DCM. This was added to a stirringmixture of purpurin-18 methyl ester (50 mg, 0.0865 mmol) in10 mL of dry DCM. The reaction was stirred for 39 h under anitrogen atmosphere with UV−vis monitoring (a shift from 700nm to 665 nm). The reaction was then treated withdiazomethane/KOH/MeOH until a red shift to 707 nmindicated the reaction was complete. A preparation scale TLCseparation was done with a 5% MeOH/DCM solvent system,and the desired band was scratched off, and the desired productwas isolated by following the method as discussed above toyield compound 5a (60 mg, 86% yield). UV−vis λmax (inCH2Cl2): 365.1 nm, 419 nm, 549 nm, 707 nm. 1H NMR (400MHz, 3 mg/1 mL CDCl3, δ ppm) 9.58, 9.32, and 8.56 (each s,1H for 5 H, 10H and 20H), 7.88 (dd, J = 8.0, 10.4 Hz 1H,31CHCH2); 6.24 (d, J = 16.0 Hz, 1H, trans 32CHCH2);6.14 (d, J = 11.6, 1H cis 32CHCH2);5.32 (m, 1H for 17H);5.10 (br s, 1H for N−CH2-CH2−(O−CH2)2-NHBoc); 4.73 (t,J = 8.0 Hz, 2H for N−CH2-CH2−(O−CH2)2-NHBoc); 4.34(q, 1H for 18H); 2.70, 2.40, 2.35, and 1.98 (each m 1H, 2 ×171H and 2 × 172H); −0.05 and −0.15 (br s, 1H for 2NH).EIMS (m/z): 809 (M + H).4.5. N-Glycol Purpurinimide Methyl Ester (5). Purpur-

in-18-N-Boc-glycol-imide (5a) was treated with a 4:1 ratio ofTFA/Dry DCM and stirred for 2 h. The reaction wasconcentrated by evaporating under vacuum. The crude mixture

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXJ

was purified on an alumina column with 5% MeOH/DCM asthe solvent system. The resultant compound after the standardworkup afforded dark purple crystals of 5 (54 mg, 89% yield).UV−vis λmax (in CH2Cl2): 365 nm (ε = 4.5 × 104), 419 nm (ε= 1.18 × 105), 551 nm (ε = 2.1 × 105), 707 nm (ε = 4.30 ×104). 1H NMR (400 MHz, 3 mg/1 mL CDCl3, δ ppm) 9.58,9.34, and 8.56 (each s, 1H for 5H, 10H and 20H); 7.90 (dd, J =11.6, 11.6 Hz, 1H for 31CHCH2); 6.30 (d, J = 16.4 Hz, 1Hfor trans -32CHCH2); 6.15(d, J = 10.4 Hz, 1H for trans -32CHCH2); 5.33 (m, 1H for 17H); 4.73 (t, J = 1.6 Hz, 2Hfor N−CH2-CH2−(O−CH2)2-NH2); 4.63 (q, 1H for18H);4.07 (m, 2H for N−CH2-CH2-(O−CH2)2-NH2); 3.85 (m, 2Hfor N-(O−CH2)2-CH2-CH2−NH2); 3.79(s, 3H for 12CH3);3.70 (t, J = 36 Hz, N-(CH2)2−O-CH2-CH2−O-(CH2)2NH2);3.62 (q, 2H for 81CH2); 3.56 (s, 3H for 172CO2CH3) ; 3.51(m, 2H for N-(CH2)2−O-CH2-CH2−O-(CH2)2NH2); 3.34 (s,3H for 18 CH3); 3.14 (s, 3H for 7CH3); 2.80, 2.70, 2.40, and1.99 (each m 1H, 2 × 171H and 2 × 172H); 1.95 (m, 2H for N-(O−CH2)2-CH2−CH2-NH2); 1.86 (broad s, 2H for NH2);1.74 (d, J = 7.2 Hz, 3H for 18CH3); 1.65 (t, J = 7.6 Hz, 3H for82 CH3); 0.059 and −0.146 (broad s, 1H for 2NH). EIMS (m/z): 709 (M + H). Elemental Anal. Calcd. for C40H48N6O6: C,67.78; H, 6.83; N, 11.86. Found: C, 67.59; H, 6.60; N, 10.90.4.6. Cyanine Dye (6). To a dry round bottomed flask

containing IR-820 (342 mg, 0.402 mmol) stirring in anhydrousDMF was added 4-carboxythiophenol (186.2 mg, 1.20 mmol),and the mixture was stirred overnight under a nitrogenatmosphere. The crude reaction product obtained after thestandard workup was purified over a silica column using a 5−25% DCM/MeOH solvent system, and a green liquid wascollected. The liquid was concentrated by evaporation undervacuum and scratched to afford dark green crystals of 6 yield(270 mg, 80%). UV−vis λmax (in MeOH) 835 nm (ε = 1.96 ×105); 1H NMR (400 MHz, 3 mg/1 mL MeOH, δ ppm) 8.87(d, 2H, J = 14 Hz), 8.15 (d, 2H, J = 14 Hz), 8.79−7.99(m, 6H),7.57−7.63 (m,4H), 7.44 (t, 2H, J = 7.2 Hz), 7.36 (d, 2H, J = 8.4Hz), 6.40 (d, 2H, J = 14 Hz), 4.27 (t, 4H, J = 7.6 Hz), 2.85−2.92(m, 8H), 1.93−2.10 (m, 10 H), 1.77 (s, 12H). EIMS (m/z): 989.4 (M+ + 2Na). HRMS: Calcd. for C53H56N2O8S3Na:967.3097. Found: 967.3127.4.7. Conjugate 7. In a dry round bottomed flask

containing cyanine dye 6 (121.62 mg, 0.1289 mmol),DMTMM (39.25 mg, 1.1 equiv) was stirred in 5 mL ofanhydrous DMF under a nitrogen atmosphere for 2 h. N-Amino purpurinimide methyl ester 3 (80 mg, 0.1289 mg) wasstirred in 5 mL of anhydrous DMF in a separate roundbottomed flask and was added to the flask containing thecyanine dye after 2 h. The reaction was stirred overnight andwas purified by a preparative silica TLC plate using a 20%MeOH/DCM solvent system. The desired band was collectedand resuspended in 5% methanol dichloromethane. It was thenfiltered, and the solvents were concentrated. The residue wascrystallized/precipitated as dark purple crystals; yield (40 mg,32%). UV−vis λmax (in MeOH): 361 nm (ε = 3.9 × 104), 412(ε = 9.6 × 104), 544 (ε = 1.7 × 104), 693 (ε = 4.5 × 104), 837(ε = 1.5 × 105). 1H NMR (400 MHz, 3 mg/1 mL MeOH, δppm) 9.72, 9.37, 8.81 (each s, 1H, for 5H, 10H and 20H); 8.78(d, J = 2.4 Hz, 2H, cyanine dye F); 8.29 (d, J = 8 Hz, 2H,cyanine dye B); 8.09 (m, 6H, cyanine dye H and D); 7.81 (d, J= 8.8 Hz, 2H, cyanine dye G); 7.64 (t, J = 12 Hz, 4H, cyaninedye C); 7.51(t, J = 8 Hz, 2H, cyanine dye A); 6.50 (d, J = 16Hz, 3H, cyanine dye E); 5.01 (m, 2H cyanine dye SO3H), 4.91(m, 1H for 17H); 4.35 (m, 4H for cyanine dye H and 1H for

18H); 3.80(br doublet, 4H, 2H for 31 CH2 and 2H for 81CH2);3.75 (d, 7H, 4H for 172CO2CH3 and 3H); 3.35 (H2O peak andhidden 3H for 12CH3 singlet peak); 3.335 and 3.225 (s, 4H, 3for 7CH3 and 1H); 2.90 (m, 5H for cyanine dye K and L); 2.50(m, 2H, 172CH3 within DMSO peak); 2.32 (m, 2H for171CH2); 2.06−1.52 (many multiplets, 41H, 12H for cyaninedye J and N, 3H for cyanine dye M, 9H for 32CH3, 18

1CH3 and82CH3, and 17 H). EIMS (m/z): 1520 (M − H). HRMS:Calcd. for C87H92N8O11S3: 1520.6048. Found: 1520.6048.4.8. Conjugate 8. In a dry round bottomed flask, cyanine

dye 6 (92.4 mg, 0.098 mmol) and DMTMM (30 mg, 0.1078mmol) were stirred in 5 mL of anhydrous DMF in a nitrogenatmosphere for 2 h. Diethlyeneaminopurpurinimide methylester 4 (70 mg, 0.098 mmol) was stirred in 5 mL of anhydrousDMF in a separate round bottomed flask and was added intothe flask containing the cyanine dye after 2 h. The reaction wasstirred overnight and was purified on a preparation scale TLCplate using a 20% MeOH/DCM solvent system. The desiredfraction was collected, and the residue obtained afterevaporating the solvent was crystallized/precipitated withDCM/hexane as dark purple crystals; yield (35 mg, 37%).UV−vis λmax (in MeOH): 365 nm (ε = 3.8 × 104), 418 nm (ε= 8.2 × 104), 549 nm (ε = 1.6 × 104), 707 nm (ε = 4.6 × 104),842 nm (ε = 1.6 × 105). 1H NMR (400 MHz, 3 mg/1 mLMeOH, δ 9.59, 9.42, and 8.86 (each s, 1H for 5H, 10H and20H); 8.61 (d, 2H for cyanine dye f, J=14 Hz); 8.16 (dd, J = 8Hz,40 Hz, 1H, 31CHCH2); 8.06 (d, 2H for cyanine dye B,J=8.4 Hz); 7.98−7.95 (m, 4H for cyanine dye H and D); 7.73−7.67(m, 4H for cyanine dye G and C); 7.47 (t, 2H for cyaninedye A, J = 7.2 Hz); 7.34 (d, 2H for cyanine dye E, J = 8.4 Hz);6.38 (t, 3H, 2H for cyanine dye E, 1H for trans 32CHCH2, J= 14 Hz); 6.20 (d, 1H for cis 32CHCH2, J = 11.6); 5.12 (d,1H for 17H, J = 1.2 Hz); 4.53−4.40(m, 3H, 1H for 18H and2H for N−CH2-CH2−N); 4.24 (broad s, 3H for cyanine dye I);3.81 (s, 3H for 12 CH3); 3.69 (m, 1H for NH); 3.60 (s, 3H for172CO2CH3); 3.39−3.31 (m, 2H for 81CH2, 3H for 2CH3 and2H for N−CH2-CH2-N, 8H for cyanine dye K and L); 3.13(s,3H for 7 CH3); 2.33−2.20 (each m, 1H for, 2 × 171H and 2 ×172H); 2.10−2.08 (each m, 8H for cyanine dye J and N); 1.925(m, 3H for 18 CH3); 1.74−1.52 (m, 15 H, 3H for 82CH3 and12 H for cyanine dye M); −0.35 and −0.41 (each br s, 1H for2NH). EIMS (m/z): 1546 (M − H). HRMS Calcd. forC89H94N8O11S3: 1546.6204. Found: 1546.6244.4.9. Conjugate 9. In a dry round bottomed flask, cyanine

dye 6 (142.7 mg, 0.151 mmol) and DMTMM (45.96 mg,0.1661 mmol) were stirred in 5 mL of anhydrous DMF in anitrogen atmosphere for 2 h. N-Glycolpurpurinimide methylester 5 (90 mg, 0.151 mmol) stirred in 5 mL of anhydrousDMF in a separate round bottomed flask was added to the flaskcontaining the cyanine dye after 2 h. The reaction was stirredovernight and was purified on a preparation scale TLC plateusing a 20% MeOH/DCM solvent system as discussed for theforegoing compound. The desired fraction was collected andconcentrated, crystalized with DCM/hexane to afford darkpurple crystals; yield (40 mg, 32%). UV−vis λmax (in MeOH):366 nm (ε = 4.5 × 104), 416 nm (ε = 1.0 × 105), 549 nm (ε =2.7 × 104), 707 nm (ε = 5.8 × 104), 842 nm (ε = 2.1 × 105). 1HNMR (400 MHz, 3 mg/1 mL MeOH): δ 9.67,9.46 and 8.88(each s, 1 H for 5H, 10H and 20H); 8.62 (d, 2H for cyaninedye f, J = 14 Hz); 8.10 (d, 4H, cyanine dye j, J = 11.6 Hz);7.85(m, 4H, cyanine dye M); 7.79 (d, 2H for cyanine dye a, J = 7.2Hz); 7.40 (d, 2H for cyanine dye b, J = 7.2 Hz); 6.39 (m, 3H,1H for 31CHCH2 and 2H for cyanine dye e); 6.19 (d, 1 H

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXK

for 32CHCH2, J = 7.6 Hz); 5.18 (m, 1H for cis -32CHCH2); 4.4 (m, 1H for 18H); 4.2 (m,1H for 17H); 3.75 (s,3Hfor 12 CH3);3.51 (s, 3H for 2 CH3); 3.30(s, 3H for172CO2CH3);3.15 (s, 3H for 7 CH3); 2.80 (m, 8H for cyaninedye i and j); 2.60, 2.55, 2.40, and 2.36 (each m, 1H for, 2 ×171H and 2 × 172H); 1.90 (m, 5H for cyanine dye c and h);1.63(s, 3H for 18 CH3); 1.50 (t, 3H for 82CH3, J=7.2 Hz);−0.35 and −0.41 (each br s, 1H for 2NH). EIMS (m/z): 1634(M − H); HRMS Calcd. for C93H103N8O13S3: 1634.6746.Found: 1634.6729.4.10. In Vivo Fluorescence Imaging. Fluorescence

imaging was performed using approved protocols in accordancewith the Guide for the Use of Laboratory Animals. A 12 bitNuance camera (CRI, Worburn, MA) was used to image theconjugates in vivo using a 782 continuous wave laser for anexcitation source. The fluorescence was collected after filteringthrough two 600 nm long pass filters in series and 800 and 830nm long pass filters. Prior to imaging, BALB/c mice bearingColon26 tumors on the shoulder with an average diameter of4−5 mm were shaved and depilated with Nair cream to removethe hair over the tumor and the surrounding skin. Conjugates 7,8, and 9 were injected via the tail vein at 0.03 μmol/kg. At 24,48, and 72 h postinjection, mice were anesthetized withketamine/xylene (100/10 mg/kg), placed in the light box, andimaged. The localization and biodistribuition of the conjugateswere determined by sacrificing 3 mice/time points and imagingthe organs. The images were processed using an imagecalculator.27

4.11. In Vivo Photosensitizing Efficacy. BALB/c mice(5 mice/group) were implanted subcutaneously in the rightshoulder with 1 × 106 Colon26 cells per 50 μL of RPMI-1640medium. When tumors grew to 4−5 mm diameter, mice wereinjected with 2.5 μmol/kg of 8 and 9 via tail vein injection.Twenty-four hours postinjection, the mice were restrainedwithin Plexiglas holders, and exposed tumors were treated withlaser light for 135 J/cm2 at 75mW/cm2. Following treatment,tumor growth was monitored daily and measured with calipersalong the length and width of the tumors. Tumor volumes werecalculated by using the formula V = LXW 2/2. Mice wereconsidered cured if no palpable tumor grew within 30 days ofbeing treated, or they were euthanized when the tumor volumereached 400 mm3.4.12. Molecular Modeling. 4.12.1. Conjugate Struc-

tures. The three-dimensional structures of purpurinimide−cyanine dye conjugates (7−9) were built with SYBYLmolecular modeling software, version 8.0 (Tripos Inc., St.Louis, MO). The purpurinimide moiety was taken fromprevious studies28 utilizing appropriate crystal structure. Thecyanine dye moiety was built from a SYBYL fragment libraryand standard geometry using the extended conformation. Bothmoieties were fully energy optimized with a semiempiricalmolecular orbital method, PM3, using the Spartan 02 software(Wave function Inc., Irvine, CA). Finally, energy optimizedmoieties were joined with specific linkers with standardgeometry using SYBYL software. Each conjugate as a wholewas again energy optimized with PM3 using Spartan software.

4.12.2. Conjugate Conformations. The PM3 energyoptimized structure for each conjugate was used as the initialconformation for limited conformational search to explore theconformational flexibility of these conjugates. Because of thesize of molecules, molecular mechanics instead of the molecularorbital method was used for this study. The stochasticconformational search was performed using Molecular

Operating Environment (MOE) software, version 2010.10,from Chemical Computing Group (Montreal, Quebec,Canada). Standard MMFFs force field and charges were usedfor the stochastic conformational search where the dihedralangles were randomly modified, followed by energy mini-mization in dihedral angles first, then energy minimization inCartesian space. During the conformational search, the chiralcenters were maintained as the original structure. No attemptwas made to perform an exhaustive conformational search dueto uncertainty in the environment of the conjugate in the cell orin an in vivo situation.4.13. Photophysical Properties. Femtosecond transient

absorption spectroscopy experiments were conducted using anultrafast source, Integra-C (Quantronix Corp.), an opticalparametric amplifier, TOPAS (Light Conversion Ltd.), and acommercially available optical detection system, Helios,provided by Ultrafast Systems, LLC. The source for thepump and probe pulses were derived from the fundamentaloutput of Integra-C (780 nm, 2 mJ/pulse and fwhm = 130 fs)at a repetition rate of 1 kHz. Seventy-five percent of thefundamental output of the laser was introduced into TOPAS,which has optical frequency mixers resulting in a tunable rangefrom 285 to 1660 nm, while the rest of the output was used forwhite light generation. Prior to generating the probecontinuum, a variable neutral density filter was inserted in thepath in order to generate a stable continuum, then the laserpulse was fed to a delay line that provides an experimental timewindow of 3.2 ns with a maximum step resolution of 7 fs. In ourexperiments, a wavelength between 350 to 450 nm of TOPASoutput, which is the fourth harmonic of signal or idler pulses,was chosen as the pump beam. As this TOPAS output consistsof not only desirable wavelength but also unnecessarywavelengths, the latter were deviated using a wedge prismwith a wedge angle of 18°. The desirable beam was irradiated atthe sample cell with a spot size of 1 mm diameter where it wasmerged with the white probe pulse in a close angle (<10°). Theprobe beam after passing through the 2 mm sample cell wasfocused on a fiber optic cable, which was connected to a CCDspectrograph for recording the time-resolved spectra (410−1600 nm). Typically, 2500 excitation pulses were averaged for 5s to obtain the transient spectrum at a set delay time. Kinetictraces at appropriate wavelengths were assembled from thetime-resolved spectral data. All measurements were conductedat room temperature, 295 K.The fluorescence measurements were carried out with an

absolute PL quantum yield measurement system (Hamamatsuphotonics Co., Ltd., C9920-02) with excitation at 680 nm.4.14. Electrochemical Measurements. Cyclic voltam-

metry was carried out with an EG&G Model 173 potentiostat/galvanostat. A homemade three-electrode cell was used andconsisted of a glassy carbon working electrode, a platinum wirecounter electrode, and a saturated calomel reference electrode(SCE). The SCE was separated from the bulk of the solution bya fritted-glass bridge of low porosity which contained thesolvent/supporting electrolyte mixture. All potentials arereferenced to the SCE.UV−visible spectroelectrochemical experiments were per-

formed with an optically transparent platinum thin-layerelectrode of the type described in the literature.29 Potentialswere applied with an EG&G Model 173 potentiostat/galvanostat. Time-resolved UV−visible spectra were recordedwith a Hewlett-Packard Model 8453 diode array rapid scanningspectrophotometer.

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXL

Tetra-n-butylammonium perchlorate (TBAP, ≥ 99%) wasrecrystallized from ethyl alcohol and dried under vacuum at 40°C for at least one week prior to use. Dimethylsulfoxide(DMSO, ≥ 99.9%) were obtained from Sigma-AldrichChemical Co. and used without further purification.

■ ASSOCIATED CONTENT*S Supporting InformationThe NMR/Mass spectra of compounds 3−9, the HPLCprofiles of conjugates 7−9, and the experimental details ofMTT assay and in vivo reflectance spectroscopy. This materialis available free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: ravindra.pandey@roswellpark.org.

■ ACKNOWLEDGMENTSThis research was supported by Grants-in-Aid (Nos. 20108010and 23750014) and the Global COE program “GlobalEducation and Research Center for Bio-EnvironmentalChemistry” of Osaka University from the Ministry ofEducation, Culture, Sports, Science and Technology, Japanand KOSEF/MEST through WCU project (R31-2008-000-10010-0) from Korea to S.F. and The Robert WelchFoundation (Grant E-680) to K.M.K. and CA 127369 (NIH)to R.K.P.

■ REFERENCES(1) Dougherty, T. J. (1996) A brief history of clinical photodynamic

therapy development at Roswell Park Cancer Institute. J. Clin. LaserMed. Surg. 14, 219−221.(2) Pandey, R. K., Goswami, L. N., Chen, Y., Gryshuk, A., Missert, J.

R., Oseroff, A, and Dougherty, T. J. (2006) Nature: A rich source fordeveloping multifunctional agents. Tumor-imaging and photodynamictherapy. Lasers Surg. Med. 38, 445−467.(3) Celli, J. P., Bryan, Q., Spring, B. Q., Rizvi, I., Evans, C. L.,

Samkoe, K. S., Verma, S., Pogue, B. W, and Hasan, T. (2010) Imagingand photodynamic therapy: mechanisms, monitoring, and optimiza-tion. Chem. Rev. 110, 2795−2838.(4) Ethirajan, M., Saenz, C., Dobhal, M., and Pandey, R. K. (2008)

The Role of Porphyrin Chemistry in Tumor Imaging and Photo-dynamic Therapy, in Advances in Photodynamic Therapy (Hamblin, M.R., and Mroz, P., Eds.) Artech House, Boston, MA.(5) Ethirajan, M., Chen, Y., Joshi, P, and Pandey, R. K. (2011) The

role of porphyrin chemistry in tumor imaging and photodynamictherapy. Chem. Soc. Rev. 40, 340−362.(6) Dougherty, T. J., Gomer, C. J., Henderson, B. W., Jori, G., Kessel,

D., Korbelik, M., Moan, J, and Peng, Q. (1998) Photodynamic therapy.J. Natl. Cancer Inst. 90, 889−905.(7) Henderson, B. W., and Dougherty, T. J. (1992) How does

photodynamic therapy work? Photochem. Photobiol. 55, 145−157.(8) Macdonald, I., and Dougherty, T. J. (2001) Basic principles of

photodynamic therapy. J. Porphyrins Phthalocyanines 5 (105), 2001−2011.(9) Gryshuk, A., Chen, Y., Goswami, L, N., Pandey, S., Missert, J. R.,

Ohulchanskyy, T., Potter, W., Prasad, P. N., Oseroff, A, and Pandey, R.K. (2007) Structure-Activity relationship among purpurinimides andbacteriopurpurinimides: Trifluoromethyl substituent enhanced thephotosensitizing efficacy. J. Med. Chem. 50, 1754−1767.(10) Chen, Y., Gryshuk, A., Achilefu, S., Ohulchansky, T., Potter, W.,

Zhong, T., Morgan, J., Chance, B., Prasad, P. N., Henderson, B. W.,Oseroff, A, and Pandey, R. K. (2005) A novel approach to abifunctional photosensitizer for tumor imaging and phototherapy.Bioconjugate Chem. 16, 1264−1274.

(11) Massound, T. F, and Gambhir, S. S. (2003) Molecular imagingin living subjects: seeing fundamental biological process in a new light.Genes Dev. 17, 545−580.(12) Brindle, K. (2008) New approaches for imaging tumor

responses to treatment. Nat. Rev. Cancer 8, 94−107.(13) Pandey, R. K., James, N., Chen, Y, and Dobhal, M, P. (2008)

Cyanine dye based compounds for tumor imaging with and withoutphotodynamic therapy. Top. Heterocycl. Chem. 14, 41−74.(14) Jager, H. R., Taylor, M. N., Theodossy, T, and Hopper, C.

(2005) MR imaging-guided interstitial photodynamic laser therapy foradvanced head and neck tumors. Am. J. Neuroradiol. 26, 1193−1200.(15) Mishra, A., Behera, R. K., Behera, P. K., Mishra, B. K, and

Behera, G. B. (2000) Cyanines during the 1990s: A review. Chem. Rev.100, 1973−2012.(16) Ornelas, C., Lodescar, R., Durandin, A., Canary, J. W., Pennell,

R., Liebes, L. F, and Weck, M. (2011) Combining aminocyanine dyeswith polyamide dendrons: A promising strategy for imaging in thenear-infrared Region. Chem.Eur. J. 17, 3619−3629.(17) Narayanan, N, and Patonay, G. (1995) A new method for the

synthesis of Heptamethine cyanine dyes: Synthesis of a new near-infrared fluorescent labels. J. Org. Chem. 60, 2391−2395.(18) Majumdar, S. R, Majumdar, R. B., Grant, C. M, and Waggoner,

A. S (1996) Cyanine-labeling reagents: sulfobenzindocyaninesuccinimidyl esters. Bioconjugate Chem. 7, 356−362.(19) Strekowksi, L., Mason, C. J., Lee, H., Gupta, R., Sowell, J, and

Patonay, G. (2003) Synthesis of water-soluble near infrared cyaninedyes functionalized with 9 succiniimdioxyl- carbonyl group. J.Heterocycl. Chem. 40, 913−916.(20) Zheng, G., Potter, R. W., Camacho, S. H., Missert, J. R., Wang,

G., Bellnier, D. A., Henderson, B. W., Rodgers, M. A., Dougherty, T. J,and Pandey, R. K. (2001) Synthesis, photophysical properties, tumoruptake, and preliminary in vivo photosensitizing efficacy of ahomologous series of 3-(1′-Alkyloxy)ethyl-3-devinylpurpurin-18-N-alkylimides with variable lipophilicity. J. Med. Chem. 44, 1540−1559.(21) Mironov, A. F., Grin, M. A., Tsiprovskiy, A. G., Kachala, V. V.,

Karmakova, T. A., Pltyutinskaya, A. D, and Yakubovskaya, R. I. (2003)New bacteriochlorin derivatives with a fused N-aminoimide ring. J.Porphyrins Phthalocyanines 7, 725−730.(22) Pandey, S. K., Gryshuk, A. L., Sajjad, M., Zheng, X., Chen, Y.,

Abuzeid, M. M., Morgan, J., Charamisinau, I., Nabi, H. A., Oseroff, A,and Pandey, R. K. (2005) Multimodality agents for tumor imaging andPDT: A possible see and treat approach. J. Med. Chem. 48, 6286−6295.(23) Kawada, K., Yonei, T., Ueoka, H., Kiura, K., Tabata, M.,

Takigawa, N., Harada, M, and Tanimoto, M. (2002) Comparison ofchemosensitivity tests. Clonogeneic assay versus MTT assay. ActaMed. Okayama 56 3, 129−134.(24) Patterson, M. S., Schwartz, E., and Wilson, B. C. (1989)

Quantitative reflectance spectrophotometry for the noninvasivemeasurement of photosensitizer concentration in tissue duringphotodynamic therapy. SPIE 1065, 115−122.(25) Potter, W. R., Bellnier, D. A., Pandey, R., Parsons, J. R., and

Dougherty, T. J. Sensitizer pharmacokinetics by in vivo reflectionspectroscopy, unpublished results.(26) Zeng, G., Chen, Y., Intes, X., Chance, B, and Glickson, J. D.

(2004) Contrast-enhanced near IR optical imaging for subsurfacecancer detection. J. Porphyrins Pthalocyanines 8, 1106−1117.(27) http://rsb.info.nih.gov/ij/(28) Pandey, S. K., Zheng, X., Morgan, J., Missert, J. R., Liu, T. H.,

Shibata, M, Bellnier, D. A., Oseroff, A. R., Henderson, B. W.,Dougherty, T. J, and Pandey, R. K. (2007) Purpurinimidecarbohydrate conjugates: effect of the position of the carbohydratemoiety in photosensitizing efficacy. Mol. Pharm. 4, 448−64.(29) Lin, X. Q, and Kadish, K. M. (1985) Vacuum-tight thin-layer

spectroelectrochemical cell with a doublet platinum gauze workingelectrode. Anal. Chem. 57, 1498−1501.

Bioconjugate Chemistry Article

dx.doi.org/10.1021/bc200345p |Bioconjugate Chem. XXXX, XXX, XXX−XXXM